Table 4 LEVEL (Logical Explanations & Visualizations of Estimates in Linear mixed models) checklist of items for reports of multilevel data and modelling analyses

Title and abstract

1

(a) Not essential in the title, but the fact that the study is hierarchical and the analyses are multilevel must be mentioned in the abstract (S)

(b) The abstract should mention the various levels considered in the analyses and whether random intercepts only or also random slopes were modeled (N)

 Background/rationale

2

Provide rationale for the study design being hierarchical and for the analyses being multilevel (E)

 Objectives

3

Mention at what level the dependent and independent variables are taken (S)

 Study design

4

(a) Provide the multilevel diagram for the study (S)

(b) Justify level of the analyses (N)

 Population

5

(a) Provide the number of members of each level, the eligibility criteria, and the sources and methods of selection/sampling of the members (N)

(b) If a repeated measures design, provide description of methods of follow-up, and spacing of time points (N)

(c) Describe missingness patterns and imbalances in members across levels (E)

 Variables/ data structure

6

(a) Write out the multilevel model equation including the random effects – this may be provided in an Appendix (N)

(b) Mention the variables used and from what level (N)

 Study size

7

(a) Provide details of the sample size calculation, and mention relevant variance partition coefficients (VPC) or intraclass correlation coefficients (ICC) and variance inflation factors (VIF) for each level (N)

(b) Provide justification for ICCs from previous studies – literature or pilot studies (E)

 Statistical methods

8

(a) Describe all statistical methods, descriptive and inferential, detailing how the correlation in the data was dealt with (E)

(b) Mention estimation procedure utilized (e.g. restricted maximum likelihood) (S)

(c) Present variance components or VPCs/ICCs for ‘null’ model and for final model (S)

(d) Justify variables considered in the initial model and justify the ones included in the final model (N)

(e) Justify choice of random or fixed intercepts and random or fixed slopes for variables in the final model, along with correlation structure among the random effects (N)

 Participants

9

(a) Report the number of individuals from each level in the final model, since missing data may affect the original numbers (N)

(b) Present a flow diagram (S)

 Descriptive data

10

(a) Indicate number of participants with missing data for each variable of interest, by level (S)

(b) Identify the level when presenting graphs and tables (E)

(c) Adjust the variances even in descriptive univariate or bivariate analyses (N)

 Modeling results

11

(a) Present the model equation and estimates – maybe in Appendix (S)

(b) Present a summary table with estimates of fixed effects, VPCs/ICCs for null model, intermediate models (if any) and final model (N)

(c) Present model goodness of fit statistics (N)

 Other analyses

12

Report other analyses and if multilevel, provide similar information as above (S)